I.  Sex with Deception

These thought experiments involve deception and sex:

Insofar as these cases involve deception, they are at least somewhat morally bad. But it is very natural to suppose that some of these lies are much worse than others—indeed, that they are listed from most to least serious. Many people would describe the deception in Identity and Condom as so serious that it amounts to sexual assault, because that kind of deception is inconsistent with consent that counts. But in cases like Interest and Taste, many will hold that the situation, while suboptimal, is still consensual. Tom Dougherty disagrees.¹ He holds that any time one agrees to something while deceived about a “dealbreaker,” that deception prevents consent that counts. A “dealbreaker” with respect to S’s agreement to Φ is, roughly, a feature of the situation such that if S knew it, S wouldn’t agree to Φ. (We’ll discuss the exact definition of a dealbreaker in Sec. V.) Since all four of our thought experiments involve dealbreakers, Dougherty will hold that they all involve sex without consent.

We resist this conclusion. We think there are important distinctions to be drawn between some of these cases—that it’s permissible to keep some things private from one’s sex partners, even when they concern dealbreakers.² But our disagreement with Dougherty isn’t merely about how to categorize dealbreakers as more or less morally significant; we have a deeper point of departure with many of the framing assumptions of the literature. On the dealbreaker framework, the problem cases arise when one agrees to some act Φ but deception invalidates consent to Φ. We emphasize a different question. On our framework, the problem arises when someone agrees to some act Φ but then some distinct act ψ, to which one has never agreed, occurs instead. So, rather than thinking that the acts have been agreed to but in ways that don’t count, we think the acts in these cases haven’t been agreed to at all.

In other words, we think the issue is about what one agrees to, not about what kind of agreement it is, that is, whether the agreement amounts to valid consent. To categorize the cases we opened with, we shouldn’t look to ethical theorizing about the circumstances under which agreements end up being morally transformative; we should look instead to metasemantic considerations about just what kind of content one’s consenting behavior signifies agreement to.³

Before motivating these conclusions, we’ll first (Sec. II) offer some clarifications about how we are using consent language. We then (Sec. III) explain Dougherty’s view. In Section IV we raise counterexamples to it. Unsurprisingly, we think Dougherty’s view categorizes some cases of intuitively consensual sex as nonconsensual. We’ll also consider cases which demonstrate that Dougherty’s view categorizes some cases of intuitively nonconsensual sex as consensual. We then (Secs. V–VI) draw on the post-Gettier epistemology literature to raise more general worries for anyone wishing to pursue a dealbreaker framework. In Sections VII–XI, we turn to our preferred view. We will argue, with common sense and against Dougherty, that only some kinds of deception, even about dealbreakers, result in nonconsensual sex. To explain this, we will shift the emphasis from questions about requirements for genuine consent to metasemantic questions about just what one is consenting to, when one consents.

II.  Agreement and Valid Consent

We begin with three clarifications.

First, according to a standard liberal conception of consent, consent is a kind of agreement to something that would otherwise be a violation of one’s rights.⁴ If you own property, for example, you have the right against trespass—we are not permitted to enter your space. But if you waive that right by consenting to our entry, we may enter without violating your rights. In the same way, people typically have the right not to be touched sexually, but they can waive that right by consenting to sexual contact.⁵

Theorists sometimes specify that valid consent is consent that genuinely waives the right—valid consent is consent that counts. Not all agreements constitute valid consent. If you agree to something because a mugger is threatening to kill you, or because you are too intoxicated to make decisions for yourself, or because you are a toddler, your agreement is said not to constitute valid consent.⁶ The terminological conventions are not entirely clear on whether such cases involve genuine, albeit invalid, consent, or whether “consent” itself must by definition be valid. We’ll speak in the latter way: all consent is valid consent; when features like threats or intoxication invalidate consent, one may agree to something, but that agreement doesn’t constitute consent. “Invalid consent,” in our parlance, is an example of something that is not consent (just as “counterfeit money” is something that is not money).

So understood, the central question of the literature to which we’re reacting can be recast thus: when does deception invalidate consent? Dougherty thinks it always does, if it involves deception about dealbreakers; Dougherty’s opponents typically say it only does when it involves certain kinds of important factors. We will eventually argue that deception isn’t invalidating consent in any of our motivating examples—although we agree that sex happens nonconsensually in some of those cases, we will offer a different kind of explanation for why this is so.

The second early clarification we wish to make is about the relationship between consent violations and wrongdoing. We do not assume that sexual violations need always involve nonconsensual sexual activity. In saying, for instance, that the sex between George and Franklin in the Interest case is consensual, we do not mean to suggest that Franklin’s deception was morally permissible, or that George has suffered no significant violation. Feminist critiques of consent theory in sexual ethics have argued that a too-narrow focus on questions about consent has distorted theorizing about sexual violation.⁷ We leave as an open question the relationship between sexual violation and consent.

Third, we’d like to say a bit about the way we describe the content of an agreement, or of consent. Just because you (validly) consent to Φ, and Φ happens, doesn’t mean it happened with your consent. An example should dismiss any potential feeling of paradox here. Suppose we ask for permission to come into your home, and you grant it. There are various ways one could describe exactly what Φ you have consented to—one might say that you consented to our entering your home, or one might say that you consented to our inspecting your windows, or one might put it in a number of other ways. Suppose one says you consented to our entering your home. The truth of this consent ascription does not in any way imply that every possible way we might enter your home is something you’ve consented to. You might consent to our entering your home and yet have your rights violated when we enter your home in a way you have not consented to. (Maybe you said, “Yes, you may come in” at the front door, and then we come back in the middle of the night and enter without knocking.)

To an approximation, we think one can say that one has consented to Φ when one consents to a more specific way of Φ-ing. So if you consent to our coming in to inspect the windows, you have consented to our coming in, but you have not consented to our coming in in the middle of the night. It follows from this way of thinking about the logic of consent that one might consent to Φ and then have Φ occur without one’s consent. (In consenting to Φ, one doesn’t consent to every way that Φ might happen.) This is what we will eventually argue is happening in the Identity and Condom cases in Section I—the subjects do consent to have sex, but they don’t consent to the sex that occurred. This will figure into our view in important ways.

First, however, we’ll discuss extant views. Most of the literature on this question has focused on questions about whether agreement constitutes valid consent in the cases in question, not on what content is agreed to. (Some of it, we think, has also muddied that distinction in an unhelpful way.) We’ll begin with our discussion of the dominant approach, starting with Tom Dougherty.

III.  Dougherty and Dealbreakers

Dougherty thinks deception about dealbreakers always undermines consent. He thinks contrary views are motivated by the thought that some features of a sex act are more important than others. Those features that are more important are called core features, while the less important features are called peripheral features. According to the view Dougherty targets, only deception that concerns core features invalidates consent, while lies about peripheral features (like natural hair color or occupation) do not violate a person’s consent. This explains why the deception in Interest and Taste doesn’t amount to an invalidation of consent—the deception in those cases concerns peripheral features rather than core ones.

Dougherty disputes the idea that some features are objectively more important than others. He describes it as “a hangover from unacceptably moralistic views of sexuality [that] has survived into more enlightened times only because it has managed to avoid being subjected to proper critical scrutiny.”⁸ His primary target is a view with normative commitments: some features of sex are objectively important, in a way that others aren’t. In these “more enlightened times,” Dougherty rejects the idea that sex acts have objectively core features, suggesting instead that “it is up to each individual to determine which features of a sexual encounter are particularly important to her.”⁹ Those features that an individual picks out as particularly important are her dealbreakers; if those features were changed, then she would not consent to the act.

Dougherty defends his central stance via analogy. He supposes that he loathes Chihuahuas, and that his friend Aisha knows this about him. He then imagines Aisha has a Chihuahua and asks if she can keep it at Dougherty’s house—but she says the dog is a Great Dane. Thinking it’s a Great Dane, Dougherty says she can keep her dog at his house. Dougherty thinks it’s obvious that Aisha has violated his consent; this is a case where “the victim’s will is opposed to what the deceiver in fact intends. The deceiver manages to obviate this obstacle to her plan by means of deception. This deception means that the victim’s acquiescence does not count as morally valid consent.”¹⁰

According to Dougherty, deception renders consent invalid in sexual interactions in exactly the same way. If someone consents because they have been deceived—that is, they would not have consented without the deception—then their consent is invalid. This is true regardless of what the “dealbreaker” is. If someone would never agree to sleep with a philosopher, then lying to that person and telling them that you are not a philosopher renders their subsequent consent invalid. What matters is not the kind of lie; what matters is whether or not the sexual interaction takes place with the consent of both parties, and deception invalidates consent whenever that deception concerns a dealbreaker for one of the parties.

Dougherty’s focus is on deception about dealbreakers, but as Hallie Liberto has pointed out, his argument is actually more general; if successful, it shows that the mere presence of a dealbreaker invalidates consent, even if no one actively engaged in deception.¹¹

IV.  Counterexamples for Dougherty

We think, endorsing pretty mainstream intuitions, that the Interest and Taste cases from Section I are counterexamples to Dougherty’s view. But in this section we’ll introduce counterexamples of a different sort. We aim in particular to challenge Dougherty’s diagnosis that the idea that sex acts have core features is a “hangover” from more regressive conceptions of sex. These examples illustrate ways opposition to Dougherty’s view can be quite progressive:

While we don’t assume our intuitions are universal, our view is that Laura need not have violated Greg’s consent by failing to disclose that she is trans,¹³ but that Bryson has violated Evelyn’s consent by deceptively removing the condom. Dougherty’s account cannot yield either of these results. Dougherty must say, since Laura allowed Greg to remain deceived about a dealbreaker (however rooted in prejudice that dealbreaker is), that Greg’s consent was invalidated.

By contrast, Dougherty must say that, although Bryson engaged in behavior that would quite often violate a partner’s consent, in this case Evelyn’s consent is valid. This is so because while many people would consider dishonesty about condom use a dealbreaker, Evelyn does not.

Both of these strike us as serious problems for Dougherty’s framework. The problem isn’t that subjects shouldn’t be able to decide for themselves what features of a sexual encounter are important to them; we agree with Dougherty that individual sexual autonomy is an important value, and that individuals should have considerable discretion along these lines. Denying this would indeed amount to an objectionable kind of sexual moralism. But Dougherty’s version goes too far—it makes sexual violation depend on individual idiosyncrasies that have never even been mentioned.

While our judgments about these cases may be somewhat controversial, it does not strike us as at all plausible that they are the product of anything like regressive moralizing about sex. On the contrary, our stance here is the “progressive” one.

So one can reject Dougherty’s approach without embracing regressive moralism. The regressiveness arises from particular assumptions about which features it’s legitimate for sex partners to care about. If one held that whether one is married to one’s partner is always of crucial importance when considering sexual consent, but that whether one has the same aesthetic tastes isn’t, then the conception of sex is regressive. Asserting that other features are or are not important—like whether or not someone is trans—can be indicative of much more progressive conceptions of sex.

At the very least, it is plausible, if not obvious, that Laura’s actions in Passing are consensual, while Bryson’s in Indifferent Condom are not. If so, the presence or absence of dealbreakers is not the key factor in these cases.

V.  Defining Dealbreakers

Here is a more general reason to worry about the emphasis on deception and dealbreakers in explaining what goes wrong in the bad cases. (Readers already convinced to abandon this framework should feel free to skip ahead to Sec. VII.)

As Dougherty puts it, S’s agreement to Φ is marred by deception in a way inconsistent with S’s valid consent to Φ if S is deceived about a dealbreaker. A dealbreaker is a feature of Φ to which S’s will is opposed, to the degree to which it “makes a decisive difference to [S’s] decision to [Φ].”¹⁴ We understand the notion of making a decisive difference in counterfactual terms. It’s not enough, for X to be a dealbreaker for your agreement to Φ, that you wouldn’t have liked X if you’d known about it; it has to be such that you would not have agreed to Φ had you known about X.

But this way of thinking about dealbreakers faces clear counterexamples. Here, for example, is a case from Hallie Liberto: “A has sex with B, knowing there is a small chance that she might get pregnant. The sex that she has with B is sex that gets A pregnant. If she had known that the sex would involve impregnation, then A would not have agreed to have sex. However, she consents to the gamble—though not to the impregnation. In this way, A has sex that involves a feature that counts as deal-breaker, for her. Yet, A has still consented to the sex.”¹⁵

Formally speaking, the fact that this sex will result in pregnancy is a dealbreaker for A. It is a feature of the situation such that, if A had known about it, A wouldn’t have agreed to sex. It is easy to imagine further examples along these lines.¹⁶ Some sex is disappointing. Many people would decline to have sex if they knew in advance that the sex would be disappointing, but their eventual disappointment doesn’t invalidate their consent. None of this remotely suggests that any such sex is nonconsensual; it is no violation of your rights, despite the presence of an unknown dealbreaker.

To be fair, the examples just mentioned do not involve active deception about dealbreakers. Still, Liberto is right that they illustrate that dealbreakers are ill-suited to invalidate consent in general—as mentioned above, Dougherty’s arguments commit him to the claim that passive ignorance of a dealbreaker invalidates consent just as much as active deceit does. Moreover, the ease with which one can construct such cases can also point the way to a broader class of counterexamples to Dougherty’s view, even ones that do involve active deceit.

Consider a dealbreaker of this form. Suppose that if A agrees to Φ, some consequence C will follow, where C is something A doesn’t want, but where if A had fuller information, A would have no objection to C. Suppose, for example, that Janet has strong political convictions and wouldn’t agree to have sex with someone if she knew he was a registered Republican. Suppose she meets Ted, who is a registered Republican. If she knew this about Ted, Janet wouldn’t agree to have sex with him, but she remains ignorant of this dealbreaker and they have sex. So far this is quite like Dougherty’s case of someone who doesn’t want to have sex with soldiers and meets someone who, unbeknownst to her, is a solider.¹⁷ But we may complicate the story, adding details that make it (even) less plausible that there is a consent problem here, while retaining the presence of a dealbreaker.

Suppose, for example, that there is what Janet herself would recognize as a good reason for Ted to be a registered Republican. Suppose Ted shares Janet’s political beliefs and values but lives in an overwhelmingly Republican state and district, such that the most meaningful vote that he can cast is in the Republican primary, where he always votes for the most liberal candidate to represent the Republican Party. To do this, he must register as a Republican.¹⁸ Once the general election comes, he always votes for the Democrat. Janet would be perfectly happy with Ted’s party registration and voting behavior if she knew all of these details, and she would agree to have sex with him.

In fact, however, Janet knows none of this. The subject of Ted’s voting behavior hasn’t come up at all. Assume she met him at a venue that makes it reasonable for her to assume that he is not a Republican—a Democratic fundraiser, perhaps, or a trans rights rally. They meet, they flirt, they go out on a date, and they decide to have sex. Ted would rather not talk about his voting strategy, so he makes a point of not mentioning it, deliberately allowing her to assume that he is not a registered Republican.

So stipulated, here are two facts:

VI.  Dealbreakers and Defeaters

There is in fact quite a strong parallel between the dialectical situation here and the twentieth-century “defeaters” approach to the analysis of knowledge. That tradition was developed in various ways, but here is one characteristic approach to identifying under what conditions S knows that P. In addition to S’s justified true belief that P, the idea was that this belief must have no defeaters. A defeater is a fact Q such that, if S knew Q, S would not have believed, or justifiably believed, that P.¹⁹ This kind of condition was suggested as a way of accounting for what goes wrong in Gettier cases to make them fall short of knowledge. For example, suppose you look at a clock that, unbeknownst to you, stopped twelve hours ago, and you form the justified true belief that it is 8:30. Your belief has a defeater: an important fact that would make a difference to your doxastic conduct, if you knew about it. More precisely, had you known that the clock was broken, you would not have formed a belief about the time on its basis. This, it was suggested, is why your belief isn’t a case of knowledge, even though you have no idea that anything is amiss.

A “no defeaters” condition for knowledge is structurally parallel to a “no dealbreakers” condition for consent. Belief is the analogue of agreement; knowledge is the analogue for consent. (Recall that we restrict “consent” to valid consent.)

Defeater-emphasizing approaches to knowledge have not fared well. They are too restrictive, for reasons very similar to the discussion of political party dealbreakers. Unless one is maximally certain that P, there will always be true defeaters to one’s belief that P. For example, consider any proposition Q that provides misleading evidence against P. Suppose Jonathan knows that he met with Emily last Tuesday, because he remembers doing so. Jonathan’s memory about these kinds of things is quite good but not infallible, and Jonathan knows that too. A “no defeaters” clause on knowledge threatens to make this quite ordinary circumstance impossible. For there will inevitably be some fact—some surprising fact—where, if presented to Jonathan out of context, it would undermine his reasonable belief that he met Emily last Tuesday. Suppose it is a fact, for example, that multiple people in the department have confused memories about last Tuesday, incorrectly taking events that day to have occurred on Monday instead. Or suppose it is a fact that Emily did not record the planned Tuesday meeting in her calendar, but instead wrote that she’d meet him Monday. (She noticed the error and showed up at the right time, but she didn’t update the calendar.) Facts like these could, if he learned them, easily lead Jonathan to lose confidence about the date of their recent meeting. Such facts, then, would count as defeaters, so Jonathan wouldn’t know.

Indeed, there is a general recipe for constructing defeaters to knowledge for anything for which one is less than maximally certain. For any subject S and proposition P such that S might be thought to know P, consider some proposition R that is surprising to S in the sense that it is true, even though S is very confident that it is false—roughly as confident as S is in P. There is no topicality restriction for R; R can be unrelated to P, so long as it is true, and S is wrong about it, and S is about as confident that P is true as that R is false. We submit that the proposition (R ∨ ¬P) will be a defeater for S knowing that P. Notice that (R ∨ ¬P) is a fact; it is a disjunction with a true disjunct, R. Moreover, if S were to learn this disjunction, S would thereby learn that they are wrong about at least one thing they had been pretty confident about, with no grounds for picking one over the other. The rational response under the circumstances would be to suspend judgment on both. So (R ∨¬P) satisfies the definition of a defeater, and so the approach to knowledge in question implies that almost nothing can be known.

The same recipe points to a very similar class of problems for the dealbreaker approach to consent. Suppose Chet only wants to have sex with natural blondes, and he meets Shannon, and they agree to have sex with each other. Even if Shannon really is a natural blonde, it is easy, following the method above, to imagine various possible facts, obviously irrelevant to the status of Chet’s valid consent, that would satisfy the definition of a dealbreaker. For example, suppose that 50 percent of the blonde women at the party where they met are not natural blondes but instead dye their hair. If Chet knew this, he might well have suspected that Shannon dyes her hair and declined to agree to have sex with her.

These cases, like those in Section IV, refute Dougherty’s view. One might attempt to rehabilitate the dealbreaker approach by drawing a subtler connection between dealbreakers and valid consent. We can’t say that the existence of the former is by itself inconsistent with the latter, but we might add further specifications as to which kinds of dealbreakers matter for consent. For example, maybe what consent requires is that there be no undefeated dealbreakers, where a defeater for a dealbreaker is something such that if you knew it, then you wouldn’t treat that thing as a dealbreaker.

The history of post-Gettier epistemology makes us pessimistic about the prospects for such a strategy. Our sense of that vast literature is that all of the good moves have been tried there, and shown wanting, and that very parallel thoughts show what’s wrong with them in this context too.²⁰ We won’t pursue these ideas further here; we leave it to those wishing to rehabilitate the dealbreaker framework to attempt to answer the challenge. Our interest is in developing an alternative account of what is going on in the cases we began with.

VII.  The Nature of the Act?

We are back to our starting point: it seems that deception about dealbreakers sometimes, but not always, results in nonconsensual sex. Dougherty denied such a distinction, but we have argued that this is untenable. So we need to explain that distinction. And we agree with Dougherty that we should do so without regressive moralizing; it is up to individuals to determine what is important when it comes to their choosing to have sex.

To avoid this regressive moralizing, we need a way of distinguishing between “core” and “peripheral” features that doesn’t rest on moral intuitions about what is more or less important. One option is to maintain that “core” features are features that are metaphysically essential to sex, while “peripheral” features are features that are not essential. This is not the route we will pursue. Instead of locating the difference between core and peripheral features in the realm of metaphysics, we’ll locate the difference within the realm of metasemantics.

The central issue, on our view, isn’t the nature of sex; it’s which, of the many acts one could have agreed to, one actually agreed to. So, in the cases under consideration, one agrees to some act Ψ but what actually occurs is a distinct act Φ. Act Φ is nonconsensual, not because it was agreed to in a way that didn’t amount to valid consent, but because it was never agreed to at all—what was agreed to was Ψ, not Φ.

What we’d like to bring out is that there are two rather different possible routes to nonconsensual sex, even if one says yes in apparent agreement to it. One—the main focus of the dealbreaker literature—is for your agreement not to count as valid consent. If one agrees to something under coercion, or while heavily intoxicated, one has not consented in the relevant, morally transformative sense. A quite different possible explanation focuses on the content of the agreement.²¹ Even supposing that every agreement one makes amounts to valid consent, we might still ask, valid consent to what? If you consent to one thing, and something else happens, then it happened without your consent.

This can even be true when both things—the thing you agree to, and the thing that happens—are (two different ways of) having sex. If so, you may agree to have sex with someone and end up having sex with someone in a different way that you didn’t agree to. If so, that happened without your consent.

Consider the legal distinction between “fraud in the factum” and “fraud in the inducement.” The latter is when one is fraudulently motivated to do something one wouldn’t otherwise have done; the former is when one is misled about what it is one is doing. For example, if someone is convinced by a lie to sign a contract (“unless you transfer the property to me, the bank will take your house”), this is fraud in the inducement; if someone signs a contract, deceived into thinking they’re signing a Christmas card, this is fraud in the factum.

We think deception about “core” features is fraud in the factum—when one is deceived about a core feature, one is deceived about what it is one is doing. We don’t deny that fraud in the inducement can happen in sexual cases, but this is not the natural way to explain the cases we’ve considered so far.²² Evelyn consented to sex with a condom, but what really occurred was sex without a condom; Thea consented to sex with Alfred, but she really had sexual contact with Don. Likewise, in his thought experiment, Dougherty consented to having a Great Dane in his home, but his friend actually brought a Chihuahua. On our view, the issue in these cases is not that they agreed to what occurred in a way that didn’t count, but rather that they didn’t agree to what occurred at all.

We said above that one way to develop this approach would be to emphasize the metaphysics of sex acts: core features are those that are definitive of the act that took place. For example, some theorists have treated cases like our Identity as nonconsensual because adultery is a fundamentally different kind of act than marital sex.²³

Our view has something in common with this one. The nature of the wrong done to Thea in our thought experiment is that she is subjected to sexual conduct to which she has not agreed. She said yes, we may imagine, but she was agreeing to have sex with her husband, not to have sex with this other person. So understood, the issue is about what is agreed to—not about the circumstances under which one’s agreement constitutes valid consent.

However, we do not think the issue depends in central moral ways on the difference between sex and adultery, two importantly distinct kinds of acts. We reject the assumption that one must locate the distinction within the metaphysics of sex acts. Even if marital impersonation cases are explicable in these terms, not all the cases that need explaining are.²⁴

Our emphasis is less on metaphysics and more on metasemantics. One needn’t posit a joint in nature between marital sex and adultery to think that consenting to sex with one’s husband is different from consenting to sex with someone disguised as him. One can agree to A without agreeing to B, even if the distinction between A and B is metaphysically shallow. (This is why we’ve been putting “core” in scare quotes.) We’ll explain in more detail how this works in the following sections.

VIII.  Content and Consent

Don disguised himself as Alfred to have sex with Thea. Our view is that Thea consented to have sex with Alfred but did not consent to have sex with Don, so her ensuing sexual contact with Don happened without her consent. Why is this?

Imagine a version of the case where the way in which Thea consents makes explicit that she is consenting to have sex with Alfred, not anyone else. She might say, for instance, “Yes, I will have marital sex with you, my husband Alfred.” Such a speech act would either constitute or express consent to a particular content: having sex with Alfred.²⁵

Being so specific in one’s explicit consent is rather artificial, but it is helpful in pointing out the role of the individual in deciding what one is consenting to do—by specifying that she means to have sex with Alfred, Thea specifies the content of what is being consented to. This point doesn’t depend in any way on their marital status, or the idea that sex with Alfred is a fundamentally different kind of activity than sex with someone else would be. While in this version of the case Thea has consented to sex with her husband, Thea could equally, if she’d liked, have consented to extramarital sex with Don but have been deceived by her husband Alfred in a Don costume; this too would have been nonconsensual. Moreover, nothing here depends on value judgments about what kind of sex is or is not legitimate. (We could even imagine—if we do not get too hung up on being particularly realistic—a version of the case where Thea thinks and says, “You look like my husband, but whether or not things are as they appear, I will have sex with you,” which would make it explicit that sex with Don in disguise is consensual, regardless of the deception.²⁶)

This consideration reflects many of Dougherty’s motivations. He, like us, wishes to emphasize individuals’ fiat to decide for themselves what they’re agreeing to. But his dealbreaker framework is passive; it’s simply a matter of how subjects would react if apprised of various circumstances. (It does not even require that the subjects know of themselves that that would be a dealbreaker for them.) We say that what matters is what one does agree to, not what one would have agreed to in counterfactual circumstances. This, we think, is a better and fuller way to implement the importance of individuals’ sexual agency.

Ours is a different kind of response to Dougherty than others in the literature, such as those by Liberto, Tadros, Lazenby and Gabriel, and Bromwich and Millum. Like those authors, we do think there can be consensual sex, even in the presence of dealbreakers. But we do not emphasize, as they do, the conditions that are necessary for an agreement to constitute valid consent.²⁷ Instead, we allow that the agreements in all the cases in question do amount to valid consent but highlight a key semantic question: what exactly are these subjects consenting to? In our view, the content of consent is a crucial but underappreciated question in this literature.

It may strike some readers as surprising that we say that deception as severe as that involved in Identity or Condom is consistent with valid consent. Do we really want to say, one might worry, that Evelyn gave morally valid consent to sex with Bryce, even though he was lying about his intention to use a condom? On our view, this would be a true but potentially misleading way to describe the situation. Recall our discussion of the logic of consent in Section II. Saying that one consents to something doesn’t mean that one consents to every possible way that it might go. In this case, Evelyn consents to sex with Bryce with a condom—which is a way of consenting to sex with Bryce—but she does not consent (or even agree, for that matter) to unprotected sex with Bryce. Since that’s what ends up happening, her consent is violated. This isn’t because of the conditions on valid consent; it’s because of the content of her consent.

The alternative view would have it that the agreement Evelyn makes when she says yes to Bryce fails to be morally transformative—it doesn’t amount to valid consent to sex at all. But attending to counterfactual circumstances reveals problems with this alternate view. Bryce lied about his intention to wear a condom, and Evelyn agreed to penetration on this basis. What if, after lying but before penetrating Evelyn, he changed his mind and decided to wear the condom after all? Then, no violation of Evelyn’s sexual rights would have occurred—the sex they’d have had, which included a condom, is exactly the kind of sex she consented to. But this implies that she did validly consent to something. The problem wasn’t that her agreement wasn’t consent; it was that she consented to something different from what actually happened.

Bromwich and Millum also emphasize questions about the content of consent, but they do so differently than we do. On their view, questions about content can explain why consent is sometimes invalid. They draw a distinction between disclosure and understanding, where both provide necessary conditions for valid consent. The disclosure conditions are about illegitimate control over the person giving consent; by disclosing relevant facts, the person receiving consent enables autonomous agreement. (Deception about dealbreakers, they think, is a way to exercise illegitimate control.) More needs to be disclosed, they argue, than needs to be understood. They think the understanding constraint requires that the person giving consent understands only (1) that they are being asked for consent, (2) how to give or refuse consent, and, importantly for our purposes, (3) what one is consenting to.²⁸ While they do emphasize questions about the content of consent, they still do so in the service of a point about the validity of consent; on their view, when there is unsuccessful communication involving a content mismatch, there is no valid consent. As we have indicated, we think this gives the wrong result about the sexual cases that motivate our project.

Our view is a bit more similar to Joseph Fischel’s. Like Dougherty, Fischel recognizes the challenges inherent in declaring some elements of sexual interactions to be “core” and others incidental, and he wishes to give individuals more autonomy to decide for themselves what they permit. But, like us, Fischel does not think that, for example, whether one was assigned a particular sex at birth need necessarily be any of one’s sex partner’s business, and so he rejects the dealbreaker framework. His strategy, like ours, is to emphasize what is actually agreed to. But Fischel focuses only on explicit conditions.²⁹ He thinks, for instance, that if someone has sex with a trans man, assuming that he is a cis man, even if his status as trans is a dealbreaker, there is only a consent violation if, when agreeing to have sex with him, they said something like, “I will only have sex with you if you were born with a penis.” The same goes for the explicit condition “I will only have sex with you if you wear a condom,” or even, less normatively, “I will only have sex with you if you don’t wear a condom.”

Making conditions like this explicit is, in our view, much too strong a requirement. We agree that when they are made explicit, if one has sex deceived about the unmet conditions, that sex is nonconsensual. But only relatively rarely are conditions given so explicitly. Most people, for example, do not specify which individuals they agree to have sex with by name or explicit thought when consenting to sex with someone. Let’s consider a more realistic version of Identity. Thea encounters Don, disguised as Alfred, and, never considering whether this might be someone other than Alfred, says to him, “Yes, I’ll have sex with you.” We think the same thing about this version of the case as we did about the version where she was explicit about agreeing only to have sex with Alfred. Her thoughts and/or actions constituted agreement to sex with Alfred; they did not constitute agreement to sex with Don. This is a fact about representation; its explanation lies within the realm of metasemantics. We don’t have anything like a full story to offer here, but it seems obvious that it will have something important to do with convention. Thea is operating in a society in which it is widely understood that a “yes” in these circumstances is an agreement to sex with her regular partner. This isn’t because of how she would react in counterfactual circumstances; it’s because the established social conventions shape the meaning of her words and thoughts. Thus, we assign a significant role to social assumptions about normality in our framework that is absent in the dealbreaker framework.

This is why, in the absence of explicit restrictions, cases like Taste do not constitute consent violations. Caroline and Shondra live (we stipulate) in the sort of social environment where it is widely understood that “white lies” about enjoying partners’ interests and speaking well of their loved ones are accepted social lubricants, and that sexual consent is not contingent on one’s private aesthetic opinions perfectly matching one’s polite compliments. (In this respect their environment is like our own.) This is so even if Caroline’s polite lie would be a dealbreaker for Shondra if she knew about it. The same goes for George and Franklin—their social environment is not one that builds honesty about what one is looking for on dating apps into consent by default. That’s not to say that Shondra and George are powerless to insist, but they would have to be explicit. Shondra could agree to have sex only with someone who likes her daughter’s art; if so, and if Caroline deceived her and had sex with her under those circumstances, it would be a consent violation. But absent any such specification here, “yes” means “yes, even if you were just being polite when you said you liked the art.” This is why there are important differences between the kinds of deception we began with.

While social assumptions do play a role in our framework, they do not override individual decisions. Convention sets the default understanding of what a “yes” means, with respect to the details that are not specified. If someone has unconventional preferences, they are not bound by the conventional expectations—but they are impacted by them. For example, they have to be explicit about their nonconventional preferences. This is one of the many respects in which things are more difficult for people whose lives don’t fit the standard social conventions. Absent such explicit specification, conventions set the default content for what is being consented to.

This is why there are important differences between the kinds of deception we began with—if you deceive someone into sex by lying about something essential to what someone agrees to, you involve them in a sex act they have not consented to, and so their sexual rights are violated. And what they’ve agreed to, given the imprecise and incomplete explicit contents of their words and thoughts, is influenced in significant ways by convention. Lies about peripheral features, on the other hand, do not change what is agreed to in the same way, and so the subsequent sex act is still the same one to which they consented. Their sexual rights are not thereby violated.³⁰

IX.  The Role of Convention

The role for convention in our story is metasemantic—we are emphasizing the role of convention in fixing the contents that one agrees to in a given act of consent. (This is analogous to the obvious role of convention in fixing the contents that one asserts in a given assertoric declaration.) We are not saying that sexual morality is a matter of convention—we are neutral on such metaethical questions. The convention concerns just what one consents to, when one’s consenting language isn’t fully explicit about salient details.³¹

In saying it’s a conventional matter, we also do not mean to suggest that it’s always easy to identify the conventions, or that everyone’s ideas about them are correct. Conventional facts are still facts that one might do a more or less accurate job identifying. It is entirely coherent to imagine cases in which there are mistakes—even widespread systematic mistakes—about what the operative conventions are. For example, wearing sexy underwear is not typically consent to sexual activity of any kind, despite the 2018 acquittal of an Irish man accused of rape, in part on the grounds that his teenage victim was wearing thong underwear. The fact that some people in the grip of rape culture might think that something conventionally indicates consent to sex doesn’t make it so.³² (We’ll say more about bad conventions in Sec. X.)

The point is a very general one; social norms have a lot to do with establishing contents as a default. Absent special contexts, making eye contact and extending one’s right arm toward someone a few feet away is an invitation to have one’s hand briefly squeezed by their hand; it’s contingent that it’s an invitation to any physical contact at all; it’s also contingent that it’s not an invitation to sexual contact. Both facts are determined in significant part by conventional norms—both statistical norms about what is typical and higher-order reactive norms about what is accepted and what is sanctioned.

The same is true for sexual interactions. There is no way to explicitly identify every form of contact that one does or does not wish to agree to, in one’s consenting behavior. Consider an example:

Think about other kinds of contact that may accompany the kiss. Having obtained consent to kiss Laura, Ravi now has permission to put his hands on her face. This is a normal thing to do while kissing, and by default, saying yes to a request for a kiss gives permission for this kind of contact too. He does not thereby have permission to put his hands on her breasts. These are contingent, conventional matters. And they change over time, with some questions currently unsettled, or indeterminate.

Consider the condom cases we’ve been discussing. We stipulated that Bryson and Evelyn agreed to have sex, using a condom, but we didn’t say what words or actions they used. Evelyn might be explicit here—“You may put your penis in my vagina, but only if it is covered by a condom!”—but this is neither necessary nor particularly typical. Condoms can enter the communicative exchange in any of a variety of ways. For example, if Evelyn and Bryson are kissing in bed, and Evelyn says, “Do you have a condom?” and he says yes and produces one, and she thereupon consents to vaginal intercourse, she has consented to sex with a condom, not to unprotected sex. Indeed, the condom need not be mentioned at all. If she simply hands him a condom at an appropriate moment, this also is sufficient to render her subsequent consent to vaginal intercourse conditional on his wearing it.

Can stated information about one’s preferences play a role in influencing what one consents to? Some kinds of examples suggest they do: suppose that on their first date, Evelyn said to Bryson, “Safety is important to me, so I always use condoms when I have sex.” On its face, this is a fact about herself that Evelyn is making known. It is not any kind of explicit restriction on consent. (There is, at this stage in the evening, no consent to sex of any kind.) Nor does it foreclose the possibility that she might make an exception and later consent to unprotected sex. Nevertheless, her saying this to Bryson at this early stage can have an effect on just what content her later consenting behavior permits. If her subsequent sexual consenting language is ambiguous out of context, within the broader context it can clearly be seen to be implicitly restricted to sex with a condom.³⁴

There are, we think, condom-specific conventions at work here. Compare two cases. In each, Evelyn and Bryson are kissing in private, and things start getting hot and heavy. In one case, as clothing starts coming off, Evelyn hands Bryson a condom; in the other, she hands him a stormtrooper helmet. Handing him a condom, under the circumstances, may actually amount to consent to penetration. The stormtrooper helmet, of course, constitutes no such consent. Moreover, handing him the condom makes it clear that her consent is restricted to sex with a condom, even if she is verbally explicit about consenting to penetration and doesn’t mention the condom. By contrast, if she handed Bryson a stormtrooper helmet while undressing, this might constitute a request or a suggestion that he wear the helmet, but if she also signals clear consent to penetration, it is not thereby conditional on his wearing the helmet, unless she makes this explicit. (If Bryson does wear the helmet as the sex begins, but he later decides it feels uncomfortable and unilaterally removes it, while continuing to penetrate her, this is no consent violation.) This is because of the particular operative social conventions having to do with condoms and sex in Bryson and Evelyn’s society. There is, so far as we are aware, no corresponding convention regarding stormtrooper helmets.

Compare language. What you mean when you say something is, in one important sense, up to you. You can say whatever you want. But the broader world plays a significant part in shaping your meaning, in multiple ways. For one thing, words themselves get their meaning at the broader social level. (It is, e.g., a wholly conventional matter that “no” means no!) For another, the availability of hermeneutical resources like words in your society makes it much easier to express some thoughts than others.³⁵ The same is true of sexual consent. If one agrees to have sex with someone, absent further specification, certain assumptions “come standard.” For example, the consent is understood by default not to specify whether you told a polite lie earlier in the evening, but it is by default understood to specify sex with a condom that remains in place until penetration ends. It does not specify whether one’s partner was born in Idaho, even if one has a strange quirk that makes them unwilling to have sex with anyone who was.

X.  Bad Conventions

One might worry that our reliance on convention means we’ll inherit regressive moralizing about sex. After all, it is undeniable that regressively moralized sexual conventions exist. If convention sets the default content consented to, regardless of whether that convention is a good convention or a bad convention, have we left the door open to regressive moralizing?

We think we are able to avoid regressive moralizing, despite granting such a central role to convention. Regressive moralizing is problematic because it is incompatible with the importance of sexual autonomy: that people ought to be able to choose for themselves what matters to them when deciding whether to have sex. Our view does not have this problem. Dougherty’s moralistic target is a view that tells people what kinds of considerations they ought to care about. Our view clearly has no such features—it is silent on what sexual priorities people ought to have.

But this silence, one might worry, leads to another problem. Consider a trans-exclusive convention (i.e., a convention such that an unspecified yes is only agreement to sex with a cis person). If such a convention is permitted to set the default content consented to, then trans people will have an obligation to disclose that they are trans before having sex. In such a context, for consent to sex to be consent to sex with a trans person, such content would have to be made explicit. This would typically require the trans person to reveal that they are trans. Such an obligation is unjust. Disclosing that you are trans can be risky and makes you vulnerable to transphobic violence. Moreover, it’s unfair; cis people have no corresponding obligation to disclose what genitals they were born with.

In order to avoid imposing such an obligation on trans people, the objection goes, we need to have at least some moralized commitments about what matters when deciding to have sex: specifically, in this case, that whether or not someone is trans does not matter, regardless of the operating conventions or someone’s explicitly specified consent. Call this view progressive moralism. The progressive moralist maintains that what was wrong with regressive moralism was not that it infringed on people’s sexual autonomy, but that it did so in the wrong ways. On this view, there are limitations on the way you can legitimately specify the contours of your consent.

We do think people should be able to choose whether to have sex according to their values.³⁶ So progressive moralism, along with regressive moralism, objectionably undermines sexual autonomy. However, we agree with the progressive moralist that the obligation generated by trans-exclusionary conventions is unjust. We do not consider this an objection to our view, because we do not think the solution to transphobia lies within consent theory. On the contrary, we consider it a theoretical benefit of our view that we can explain part of what makes trans-exclusionary conventions so problematic—they impose unjust and risky obligations on trans people.

Moreover, we think our framework can make sense of what feels so complicated about this kind of case. Consider again Passing from Section IV, about Laura, who doesn’t tell Greg that she is trans before having sex with him. We indicated that we do not consider this to be a violation of Greg’s consent, even if it is a dealbreaker for him, although we noted at the time that judgments about this question are controversial. While we do not agree with them, we acknowledge that some people would consider disclosure on Laura’s part of her status as trans to be necessary for valid consent. It’s a question that genuinely divides some people at present.

This social ambivalence, we suggest, derives from a genuine indeterminacy in the present social norms. In the absence of particular signals that one is only willing to have sex with someone who’s gender matches their sex assigned at birth, it may not be wholly determinate whether unwitting sex with a trans person is consistent with what one has consented to. There may be an insufficiently robust set of social background norms to adjudicate the question. This is highly contingent, and also highly local—we have already signaled that, in the present authors’ own social circles, we do not think there is a norm to the effect that consent is by default conditional on all parties being cisgender. In these social environments, having been assigned a different gender at birth is typically classed as a relatively incidental historical feature, like having been born in Idaho, for the purpose of consent. Just as, in saying yes to a woman, one by default means “yes, even if you were born in Idaho,” one also by default means “yes, even if you were designated male when you were born.”

But we also recognize that there are more trans-exclusionary social environments, where the norms may be more cisnormative. We do not condone such norms—in our view they are inconsistent with a society affirming of the diversity of human identities. But we do think they exist. This is why the idea of so-called “gender fraud” prosecutions is plausible to some.³⁷ In such cases, where the norms really are up for negotiation, disputes about whether such cases are consensual are not merely factual or descriptive disputes; they are disputes about the conventions of consent. This, we think, is part of the explanation for why they feel so politically weighty: they are. This is not because they are disputes about the metaphysics of sex acts—it’s no part of our view that sex with a trans person is or may be different in a deep metaphysical sense from sex with a cis person—but because they are disputes about the conventions of default agreement: does or might an unqualified yes encode trans-exclusionary exceptions? If so, those conventions will set the default content of what is consented to.³⁸

We would like to emphasize that we condemn conventions that encode trans-exclusionary practices—but just because we condemn these conventions doesn’t mean they don’t have the same effect on meaning as good or neutral conventions. In fact, we think this implication serves to underscore the importance of changing these problematic conventions.

XI.  Deception

Our central questions were motivated by cases involving deception. But on the view we have defended, deception does not after all play a central role. Deception entered the picture as a hypothesis as to why, in some cases, agreement doesn’t constitute consent. What we have suggested is that in at least many of the cases at issue, there simply was no agreement to engage in the sexual conduct in question, so we do not need deception to explain why it doesn’t amount to consent.

The central question in the dealbreaker framework was, under what circumstances does S’s agreement to Φ fail to amount to valid consent to Φ? We have argued that there are serious challenges to many of the central strategies for answering that question. Our view is different. Rather than focusing on the circumstances under which consent is invalidated, we have suggested that the place to look to explain what is going on is in the specification of the content consented to. In other words, instead of asking when S’s agreement to Φ falls short of valid consent to Φ, we have suggested that in the cases at issue S may have given valid consent to some Ψ but have not agreed to Φ. So deception is merely contingently connected to the problems at hand.

Recall the Condom case from Section I. While there was obviously deception in this case, and we agree that Evelyn has not given valid consent to what happens, we do not think the latter depends essentially on the former. The dealbreaker framework has it that Evelyn has agreed to their interaction in a way that falls short of valid consent; we suggest that the problem is simpler: she simply never agreed to unprotected sex. We hold that she gave perfectly valid consent to sex with a condom, but gave no agreement whatsoever—not even deal-broken invalid consent—to sex without a condom.

Once we spell it out this way, the deception becomes inessential. It is part of the causal process by which Bryson put himself in a position to violate Evelyn the way he does, but it is, on our view, not an essential part of the violation. We think this gets the intuitions right about nearby cases. We can imagine a version of the case, for example, where there is no deception, but where the violation is the same. For example:

Bryson behaves externally the same way in Condom* as he did in Condom; the difference is in his private mental state at the time they agree to have sex with a condom. In our original version, he was deceiving Evelyn; in the new version, he was entirely sincere. In both cases, we submit, he subjects Evelyn to nonconsensual sexual contact, so the violation of her consent does not depend on his deception. (We don’t have to assume that he makes a point of hiding his condom removal; perhaps Evelyn just doesn’t notice.) We don’t see that removing the deception from Bryson’s statements or actions makes any significant moral difference to the sexual harm Evelyn suffers. If there is a moral harm attached to the lie in the first two versions of the case, it is not one that the later sexual harm depends on. The problem is that Evelyn is subjected to sexual activity—penetration without a condom—to which she has never even agreed, let alone consented.

XII.  Conclusion

There are certainly many important questions still to be answered, but we hope that we have succeeded in shifting the focus to the right set of questions and offered some preliminary analysis of how to handle them. We reject the dealbreaker proposal, although we retain some of the motivation behind it. More broadly, we reject the project of identifying the wrong in cases like those given in Section I in terms of deception. While deception and other deficiencies might invalidate consent in some circumstances, that is not how to describe the problem in these cases. The problem is that things happened that weren’t agreed to. The central theoretical question isn’t an ethical question about undermining consent, or a metaphysical question about the nature of sex; it’s a semantic question about what exactly is being agreed to by a given act of consent. Convention has a role to play in this question, but not one that supersedes individual autonomy.